Bug 59436 - [845g] Invalid PTE entries for BIOS fb during KMS takeover
Summary: [845g] Invalid PTE entries for BIOS fb during KMS takeover
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86 (IA32) Linux (All)
: medium major
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-01-15 20:25 UTC by Sandy Beach
Modified: 2017-07-24 22:59 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
syslog kernel commandline arguments loglevel=7 drm.debug=9 (487.01 KB, text/plain)
2013-01-15 20:25 UTC, Sandy Beach
no flags Details
dmesg parses easyier then syslog (84.18 KB, text/plain)
2013-01-15 20:28 UTC, Sandy Beach
no flags Details
l/proc/sys/devbug/dri/0/i915_error_state (1.97 MB, text/plain)
2013-01-15 20:36 UTC, Sandy Beach
no flags Details
/usr/bin/intel_reg_dumper (9.75 KB, text/plain)
2013-01-15 20:37 UTC, Sandy Beach
no flags Details
/sys/devices/pci0000:00/0000:00:02.0/rom (48.00 KB, text/plain)
2013-01-15 20:40 UTC, Sandy Beach
no flags Details
xserver log (33.42 KB, text/plain)
2013-01-15 20:41 UTC, Sandy Beach
no flags Details

Description Sandy Beach 2013-01-15 20:25:16 UTC
Created attachment 73116 [details]
syslog   kernel commandline arguments loglevel=7 drm.debug=9

This is a default debian testing installation using one system partition
and a swap partition. more info about the default 3.2.0-4-686-pae behavior is available if needed. 

Booting to 3.8.0-rc2+
when the xserver starts GDM3 displays only a backlite screen (basically a blank screen) with a visible functional mouse pointer.

syslog line 1201 - 1205

Jan 14 17:04:04 grg13 kernel: [   23.945802] [drm:drm_ioctl], pid=2260, cmd=0x6458, nr=0x58, dev 0xe200, auth=1
Jan 14 17:04:04 grg13 kernel: [   23.953592] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state
Jan 14 17:04:04 grg13 kernel: [   23.957573] i915: render error detected, EIR: 0x00000010
Jan 14 17:04:04 grg13 kernel: [   23.957573] [drm:i915_report_and_clear_eir] *ERROR* EIR stuck: 0x00000010, masking
Jan 14 17:04:04 grg13 kernel: [   23.957573] i915: render error detected, EIR: 0x00000010

Hardware:

Intel D845GRG  - Documentation urls see (1)
i845g chipset
2.26 GHz P4 processor
Kingston ddr pc2100  512 MB
3.8.0-rc2+ 
i686
 
more hardware info available if needed.

Installation media: http://cdimage.debian.org/cdimage/wheezy_di_beta4/multi-arch/jigdo-cd/debian-wheezy-DI-b4-amd64-i386-netinst.jigdo

Installed Package:

i   linux-image-3.8.0-rc2+              Version: 3.8.0-rc2+-1
i   linux-headers-3.8.0-rc2+            Version: 3.8.0-rc2+-1
i   linux-firmware-imag                 Version: 3.8.0-rc2+-1
i A linux-libc-dev                      Version: 3.8.0-rc2+-1
ii  xserver-xorg                         1:7.7+1
ii  xserver-xorg-core                    2:1.12.4-4
ii  xserver-xorg-input-evdev             1:2.7.0-1+b1
ii  xserver-xorg-video-fbdev             1:0.4.2-4+b3
ii  xserver-xorg-video-vesa              1:2.3.1-1+b1
ii  xserver-xorg-video-intel             2:2.19.0-6
ii  libdrm-intel1:i386                   2.4.33-3
ii  libdrm2:i386                         2.4.33-3
ii  mesa-utils                           8.0.1-2+b3
ii  libgl1-mesa-dri:i386                 8.0.5-3
ii  libgl1-mesa-glx:i386                 8.0.5-3

downstream asked me to elevate the severity. irc

(1)

D845GRG Product Home - http://www.intel.com/p/en_US/support/highlights/dsktpboards/d845grg
D845GRG Technical Product Specification: April 2002 - http://downloadmirror.intel.com/15328/eng/D845G_Boards_TechProdSpec.pdf
D865GRG Product Supplemental Document: April 2002 - http://downloadmirror.intel.com/15328/eng/D845GRG_ProductSupplement.pdf
D845GRG Technical Product Specificatin Update: November 2003 - http://downloadmirror.intel.com/15328/eng/D845G_Boards_SpecUpdat
e07.pdf
D865GRG Product Supplemental Document update: November 2003 - http://downloadmirror.intel.com/15328/eng/D845GRG_SpecUpdate07.pd
f
D845GRG BIOS Setting by Menu - http://download.intel.com/support/motherboards/desktop/sb/biosglossarybymenu_v13.pdf
D845GRG BIOS Setting alphabetical - http://download.intel.com/support/motherboards/desktop/sb/biosglossaryalpha_v13.pdf

D845GRG Users Guide and Specifications) - http://downloadmirror.intel.com/15328/
eng/D845GRG_D845GBV_ProductGuide_English.pdf

attached: syslog

other attachments: dmesg i915_error_state, intel_reg_dump, vbios.dump, Xorg.0.log
Comment 1 Sandy Beach 2013-01-15 20:28:35 UTC
Created attachment 73117 [details]
dmesg    parses easyier then  syslog
Comment 2 Sandy Beach 2013-01-15 20:36:03 UTC
Created attachment 73118 [details]
l/proc/sys/devbug/dri/0/i915_error_state
Comment 3 Sandy Beach 2013-01-15 20:37:19 UTC
Created attachment 73120 [details]
/usr/bin/intel_reg_dumper
Comment 4 Sandy Beach 2013-01-15 20:40:32 UTC
Created attachment 73121 [details]
/sys/devices/pci0000:00/0000:00:02.0/rom
Comment 5 Sandy Beach 2013-01-15 20:41:10 UTC
Created attachment 73122 [details]
xserver log
Comment 6 Sandy Beach 2013-01-23 12:23:42 UTC
Booting with debian 3.7-trunk-686-pae to runlevel 4 and the xserver ran gnome for 48+ hours, This was the first time I've had gnome 3 running on this machine. Odd, I removed gdm3 from runlevel 4 so I could build 3,8rc4 without having to shutdown gdm3.
While in gnome, screen refresh rate / rendering was extremely slow, other than that it looked like a normal gnome in fallback mode.
Booting to 3.8rc2 has not brought up a visible screen gdm3 yet.
I saved a new set of logs while gnome was working.
Planning to build 3.8rc4 and test.
Comment 7 Sandy Beach 2013-01-24 21:22:40 UTC
 3,7-trunk-686-pae boots to working gnoe in failback mode attaching a new
set of logs.
dmesg
i915_error_state 
intel_reg_dump 
syslog 
vbios.dump 
Xorg.0.log

booting 3.8-rc4
there are still no draws to the screen.
Again blank - backlite screen with  working mouse cursor.
Comment 8 Sandy Beach 2013-04-15 19:01:41 UTC
I zeroed out the disk and reinstalled using:
http://cdimage.debian.org/cdimage/weekly-builds/multi-arch/jigdo-cd/debian-testing-amd64-i386-netinst.jigdo
from weekly builds.
The install looked solid.
Will reproduce logs at your request. thanks 

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=698197


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.